This proposal represents a systematic investigation of memory-induced performance costs in amnesia. Until recently, investigations of spared memory capacities in amnesia have focused on the facilitatory effects that an experience can have on subsequent task performance. Recent studies in normal cognition, however, have highlighted the fact that experience-induced facilitation comes at a price: Enhanced processing of one stimulus is often accompanied by impaired processing of another. Surprisingly little is known about whether and under what conditions amnesic patients show the memory-induced performance costs that have been observed in normal cognition. The present studies are guided by a framework that distinguishes among three kinds of memory-induced performance costs: 1) those that are the product of explicit memory processes operating in nominally implicit memory tasks; 2) those that are the product of implicit memory processes and that may be opposed by explicit memory processes; and 3) those that are the consequence of inhibitory control processes that enhance stimulus selection, and that operate independently of explicit memory. We predict that performance costs in amnesia will be impaired or absent in the first instance, normal or greater-than-normal in the second instance, and normal in medial temporal-lobe amnesia but impaired in Korsakoff amnesia in the third instance. A better understanding of the conditions under which amnesic patients do or do not show memory-induced performance costs will elucidate the scope and nature of spared memory capacities in amnesia, and will constrain theories about the mechanisms underlying performance costs in normal cognition.